|Publication number||US4912634 A|
|Application number||US 07/072,709|
|Publication date||Mar 27, 1990|
|Filing date||Jul 13, 1987|
|Priority date||Oct 15, 1986|
|Publication number||07072709, 072709, US 4912634 A, US 4912634A, US-A-4912634, US4912634 A, US4912634A|
|Inventors||Tadayoshi Nakano, Toyohiko Yoshida|
|Original Assignee||Mitsubishi Denki K. K.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (12), Non-Patent Citations (1), Referenced by (5), Classifications (7), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The present invention relates to data processing system equipped with an instruction prefetch device.
2. Description of the Prior Art
FIG. 4 shows the instruction prefetch section of a conventional data processing system or computer having a plurality of types of instructions and an instruction prefetch queue. It consists of a main memory 1, an instruction prefetch queue 2, a processing section 3 for processing the instruction base portion, a processing section 4 for processing the instruction extension portion, a pointer PI for instruction prefetch queue input, and a pointer PO for instruction prefetch queue output.
The operation will be described with reference to FIG. 5. Suppose that instruction types include a type 1 consisting of an operation code (one word length) and a literal (one word length) and a type 2 consisting of only an operation code (one word length). The word herein used has an appropriate length for expressing an operation code and a literal code. The literal code means displacement value, immediate value, absolute value (address), etc. The operation code and literal code are processed in the processing sections 3 and 4, respectively. The process is carried out in phases. The phase herein used means the time required to fetch one word from the instruction prefetch queue 2 or advance the pointer. In order to make the description clearer, respective phases will be given numbers, such as PH 1, PH 2, . . . . The same reference numerals or characters denote identical parts or equivalents in the drawings throughout the description.
In the PH 1, a word of an instruction is fetched into the processing section 3. In the PH 2, the fetched instruction is analyzed and the pointer PO's value is incremented by one. In the PH 2, the type of the fetched instruction is also determined. If it is a type 1, one word is fetched to the processing section 4 in the PH 3. While the processor 4 processes it, the pointer PO's value is simultaneously incremented by one in the PH 4. In the PH 5, a word of another instruction is fetched to the processing section 3. In the PH 6, the fetched instruction is analyzed while the pointer PO's value is incremented by one. If the instruction is determined to be a type 2, a word of another instruction is fetched to the processing section 3 instead of the processing section 4 in the PH 7. In this way, the types 1 and 2 of instructions are processed.
As has been described above, in the conventional computer equipped with an instruction prefetch queue, either the processing section 3 or processing section 4 can operate at a time, thus requiring four phases to process a type-1 instruction and two phases to process a type-2 instruction and has poor processing efficiency.
It is a primary object of the invention to provide a data processing system having high process speed.
According to the invention there is provided a data processing system having a main memory, which comprises instruction prefetch queue for fetching instructions from said main memory; processing unit for interpreting said instructions from said instruction prefetch queue; and selection unit for selecting M continuously cyclic control signals from N control signals, where M is smaller than N, and outputting them for said processing means to fetch said instructions from said instruction prefetch queue.
Other objects, features, and advantages of the invention will be apparent from the following description when taken in conjunction with the accompanying drawings.
FIG. 1 is a block diagram of the instruction fetch section of a data processing system embodying the present invention.
FIGS. 2(a), (b), and (c) show sets of pointer combinations and their cyclic changes.
FIG. 3 is a chart for illustrating the operation of an instruction prefetch queue according to the invention.
FIG. 4 is a block diagram of the instruction fetch section of a conventional computer.
FIG. 5 is a chart for illustrating the operation of an instruction prefetch queue according to the prior art.
Referring now to FIG. 1 there is shown the instruction fetch section of a data processing system or computer according to the present invention. The same reference numerals as those of FIG. 4 denote identical parts. A three-input, two-output multiplexer 5 serves as a selection circuit for selecting two output pointers from three output pointers PO1, PO2, and PO3. The first, second and third pointers PO1, PO2, and PO3 are used for the instruction prefetch queue outputs, respectively, which are to be input to the multiplexer 5. SPO1 and SPO2 are the fourth and fifth pointers useful for instruction prefetch queue outputs selected as output control signals. In this embodiment, a plurality of types of instructions may be processed so that the prefetch section is adapted to store a plurality of types of instructions. For this reason, there are provided a pair of processing sections; a first processing section for processing the basic portion of an instruction and a second processing section for processing the extended portion of an instruction. A plurality of second processing sections may be provided. The selection circuit 5 is capable of selecting successive M control lines out of N control lines (M<N). The Nth control line is wrapped around to the first control line, and M is a given number of successive signal lines.
The operation will be described assuming that the instruction types are the same as those of the conventional and the phases are given numbers for clearer description. First of all, the operation of the multiplexer 5 in FIG. 1 will be described. The multiplexer 5 receives pointers PO1, PO2, and PO3 and outputs a set of pointers PO1 and PO2, PO2 and PO3, or PO3 and PO1. Let the set of SPO1=PO1 and SPO2=PO2 be a mode A, the set of SPO1=PO2 and SPO2=PO3 a mode B, and the set of SPO1=PO3 and SPO2=PO1 a mode C (see FIG. 2(a)). When the instruction type is determined, the mode is switched in the direction of (C)→(B)→(A)→(C) for the type 1 (FIG. 2(c)) and (A)→(B)→(C)→(A) for the type 2 (FIG. 2(b)).
The operation of the instruction fetch section will be described with reference to FIG. 3. Suppose that the multiplexer 5 is in the mode A. In the PH 1, the processing sections 3 and 4 each fetch a word of the instruction in the prefetch queue 2 specified by the pointers SPO1 (=PO1) and SPO2 (=PO2), respectively. In the PH 2, the pointer PO1, which has been connected to the pointer SPO1 in the PH 1, is increased by three while the processing section 3 analyzes the instruction and the processing section 4 processes the instruction. At this point, the instruction type is determined. Suppose now that it is a type 1. Then, the mode of the multiplexer 5 is switched from (A) to (C) and the pointers SPO1 and SPO2 are connected to the pointers PO3 and PO1, respectively.
In the PH 3, the pointer PO2, which is connected to neither pointer SPO1 nor SPO2, is increased by three while the processing sections 3 and 4 fetch a word of the instruction specified by the pointers SPO1 (=PO3) and SPO2 (=PO1), respectively. In the PH 4, the pointer PO3 connected to the pointer SPO1 is increased by three while the processing section 3 analyzes the instruction and the processing section 4 processes the instruction. At this point, the instruction type is determined. Suppose now that it is a type 2 this time. Then, the mode of the multiplexer 5 is switched from (C) to (A) and the pointers SPO1 and SPO2 are connected to the pointers PO1, and PO2, respectively, while the contents of the processing section 4 are cancelled. In the PH 5, the processing sections 3 and 4 each fetch a word of the instruction specified by the pointers SPO1 (=PO1) and SPO2 (=PO2), respectively. In this way, the processing sections 3 and 4 operate in parallel, thus reducing the processing time. In this embodiment there are three modes A, B, and C, but the number of modes increases as M and N increase.
As has been described above, according to the invention there is provided a data processing system which comprises instruction prefetch queue for fetching instructions from said main memory; processing unit for interpreting said instructions from said instruction prefetch queue; and selection unit for selecting a predetermined set of control signals from a plurality of control signals input and output for said processing unit to fetch said instruction from said instruction prefetch queue, so that the prefetch output may be controlled efficiently, thus increasing the process efficiency.
While a preferred embodiment of the invention has been described above, such description is for illustrative purpose only, and it is to be understood that changes and variations may be made without departing from the spirit and scope of the invention as defined in the following claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4110822 *||Jul 11, 1977||Aug 29, 1978||Honeywell Information Systems, Inc.||Instruction look ahead having prefetch concurrency and pipeline features|
|US4360868 *||Nov 5, 1979||Nov 23, 1982||Data General Corporation||Instruction prefetch means having first and second register for storing and providing a current PC while generating a next PC|
|US4551799 *||Feb 28, 1983||Nov 5, 1985||Honeywell Information Systems Inc.||Verification of real page numbers of stack stored prefetched instructions from instruction cache|
|US4561052 *||Sep 7, 1982||Dec 24, 1985||Tokyo Shibaura Denki Kabushiki Kaisha||Instruction prefetch system|
|US4691277 *||Oct 24, 1984||Sep 1, 1987||International Business Machines Corp.||Small instruction cache using branch target table to effect instruction prefetch|
|US4709324 *||Nov 27, 1985||Nov 24, 1987||Motorola, Inc.||Data processor control unit having an interrupt service using instruction prefetch redirection|
|US4714994 *||Apr 30, 1985||Dec 22, 1987||International Business Machines Corp.||Instruction prefetch buffer control|
|US4722047 *||Aug 29, 1985||Jan 26, 1988||Ncr Corporation||Prefetch circuit and associated method for operation with a virtual command emulator|
|US4747045 *||Jun 26, 1985||May 24, 1988||Nec Corporation||Information processing apparatus having an instruction prefetch circuit|
|US4763245 *||Oct 30, 1985||Aug 9, 1988||International Business Machines Corporation||Branch prediction mechanism in which a branch history table is updated using an operand sensitive branch table|
|US4791557 *||Jul 31, 1985||Dec 13, 1988||Wang Laboratories, Inc.||Apparatus and method for monitoring and controlling the prefetching of instructions by an information processing system|
|US4847748 *||Mar 25, 1987||Jul 11, 1989||Nec Corporation||Virtual memory arrangement data processing system with decoding and execution of prefetched instructions in parallel|
|1||*||The Bipolar Digital Integrated circuits data Book for Design Engineers, Part 2, TTL & Bipolar Memory, 1982, Nippon Texas Instruments.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5150468 *||Jun 30, 1989||Sep 22, 1992||Bull Hn Information Systems Inc.||State controlled instruction logic management apparatus included in a pipelined processing unit|
|US5317701 *||Jan 19, 1993||May 31, 1994||Motorola, Inc.||Method for refilling instruction queue by reading predetermined number of instruction words comprising one or more instructions and determining the actual number of instruction words used|
|US5774712 *||Dec 19, 1996||Jun 30, 1998||International Business Machines Corporation||Instruction dispatch unit and method for mapping a sending order of operations to a receiving order|
|US7767574 *||Mar 30, 2007||Aug 3, 2010||Kabushiki Kaisha Mikuni Kogyo||Method of forming micro metal bump|
|EP0436341A2 *||Dec 17, 1990||Jul 10, 1991||Motorola, Inc.||Sequential prefetch method for 1, 2 or 3 word instructions|
|U.S. Classification||712/207, 712/E09.055|
|Cooperative Classification||G06F9/3802, G06F9/3814|
|European Classification||G06F9/38B8, G06F9/38B|
|Jul 13, 1987||AS||Assignment|
Owner name: MITSUBISHI DENKI KABUSHIKI KAISHA, 2-2-3 MARUNOUCH
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNORS:NAKANO, TADAYOSHI;YOSHIDA, TOYOHIKO;REEL/FRAME:004745/0197
Effective date: 19870616
Owner name: MITSUBISHI DENKI KABUSHIKI KAISHA, JAPAN
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:NAKANO, TADAYOSHI;YOSHIDA, TOYOHIKO;REEL/FRAME:004745/0197
Effective date: 19870616
|Sep 13, 1993||FPAY||Fee payment|
Year of fee payment: 4
|Sep 15, 1997||FPAY||Fee payment|
Year of fee payment: 8
|Oct 16, 2001||REMI||Maintenance fee reminder mailed|
|Mar 27, 2002||LAPS||Lapse for failure to pay maintenance fees|
|May 21, 2002||FP||Expired due to failure to pay maintenance fee|
Effective date: 20020327